On Fri, 26 Dec 2003 15:17:13 +0100, Peter Larsen
wrote:

Stewart Pinkerton wrote:

Nobody seem to worry about air distortion ...

Could you give me a pointer to such a site? It's the idea of a *big*
hole in the ceiling that is caused WAF problems.

http://t-3.cc/users/audioworx/page13-12Shiva1.html

- a subpage of http://www.klone-audio.com/ - 12 large bass units firing
into a box with an estimated 12 by 12 inch opening into the listening
room.

Very impressive! Especially considering the ratio of driver area to
throat area.

For a more practical implementation of using multiple loudspeakers one
should in my opinion think differently. I would arrange a row of
loudspeakers on the loft in each side of the room and let the sound
enter the room via a long slit near the ceiling/wall junction. It would
still be possible to mount the loudspeakers vertically for longevity
(playing "downwards-sideways" into the slit) and it could be implemented
very discreetly. It will also provide a superiour coupling to the room.

A slit is certainly an option in my room, although I might have to
write my own modelling program! :-(

Peerless has a 10" sub-bass that looks about right if used in multiples.

At the moment, a pair of Adire Tempests looks to be the most likely
solution.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Fri, 26 Dec 2003 15:17:13 +0100, Peter Larsen
wrote:

Stewart Pinkerton wrote:

Nobody seem to worry about air distortion ...

Could you give me a pointer to such a site? It's the idea of a *big*
hole in the ceiling that is caused WAF problems.

http://t-3.cc/users/audioworx/page13-12Shiva1.html

- a subpage of http://www.klone-audio.com/ - 12 large bass units firing
into a box with an estimated 12 by 12 inch opening into the listening
room.

Very impressive! Especially considering the ratio of driver area to
throat area.

For a more practical implementation of using multiple loudspeakers one
should in my opinion think differently. I would arrange a row of
loudspeakers on the loft in each side of the room and let the sound
enter the room via a long slit near the ceiling/wall junction. It would
still be possible to mount the loudspeakers vertically for longevity
(playing "downwards-sideways" into the slit) and it could be implemented
very discreetly. It will also provide a superiour coupling to the room.

A slit is certainly an option in my room, although I might have to
write my own modelling program! :-(

Peerless has a 10" sub-bass that looks about right if used in multiples.

At the moment, a pair of Adire Tempests looks to be the most likely
solution.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

"Stewart Pinkerton" wrote in message
...
On Fri, 26 Dec 2003 16:35:18 -0600, "Rusty Boudreaux"
wrote:

"Stewart Pinkerton" wrote in message
...
Thanks Rusty, I did find them after trawling the Stryke site.
I've
e-mailed them for a quote, as their on-line store implies
world-wide
shipping, so we'll see what transpires. I must say that they
are the
coolest *looking* drivers around, with those big alloy cones!

I think the Tumults are more impressive but the AV15s are
certainly awe inspiring! During the first preorder (over a year
ago) the drivers came with a logo on the cone by default. You
may have to specify no logo. Stryke is essentially a one man
shop (John E. Janowitz). I think his wife just had a baby so he
may be somewhat tied up. His phone is 920-469-9198 if all else
fails. There's also a Stryke web forum.
http://forum.stryke.com/

Thanks for that, Rusty. I must say, the recent discussions on the
Stryke forum are *not* encouraging to a potential purchaser!

I appreciate the problems if he's a 'one man band' and a new dad, but
at the end of the day I'll need products, not promises............
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Keep us informed on the project and it's results.

I'm planning a project with a single Shiva in a ported box of either 3.6 or
4.8 cu ft. f3 predictions are 25 and 21 Hz, respectively.

Leaning towards the smaller box for WAF nad with room gain it should get to
at least 20 Hz in the room it will be used in.

"Stewart Pinkerton" wrote in message
...
On Fri, 26 Dec 2003 16:35:18 -0600, "Rusty Boudreaux"
wrote:

"Stewart Pinkerton" wrote in message
...
Thanks Rusty, I did find them after trawling the Stryke site.
I've
e-mailed them for a quote, as their on-line store implies
world-wide
shipping, so we'll see what transpires. I must say that they
are the
coolest *looking* drivers around, with those big alloy cones!

I think the Tumults are more impressive but the AV15s are
certainly awe inspiring! During the first preorder (over a year
ago) the drivers came with a logo on the cone by default. You
may have to specify no logo. Stryke is essentially a one man
shop (John E. Janowitz). I think his wife just had a baby so he
may be somewhat tied up. His phone is 920-469-9198 if all else
fails. There's also a Stryke web forum.
http://forum.stryke.com/

Thanks for that, Rusty. I must say, the recent discussions on the
Stryke forum are *not* encouraging to a potential purchaser!

I appreciate the problems if he's a 'one man band' and a new dad, but
at the end of the day I'll need products, not promises............
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Keep us informed on the project and it's results.

I'm planning a project with a single Shiva in a ported box of either 3.6 or
4.8 cu ft. f3 predictions are 25 and 21 Hz, respectively.

Leaning towards the smaller box for WAF nad with room gain it should get to
at least 20 Hz in the room it will be used in.

"Stewart Pinkerton" wrote in message
...
On Fri, 26 Dec 2003 16:35:18 -0600, "Rusty Boudreaux"
wrote:

"Stewart Pinkerton" wrote in message
...
Thanks Rusty, I did find them after trawling the Stryke site.
I've
e-mailed them for a quote, as their on-line store implies
world-wide
shipping, so we'll see what transpires. I must say that they
are the
coolest *looking* drivers around, with those big alloy cones!

I think the Tumults are more impressive but the AV15s are
certainly awe inspiring! During the first preorder (over a year
ago) the drivers came with a logo on the cone by default. You
may have to specify no logo. Stryke is essentially a one man
shop (John E. Janowitz). I think his wife just had a baby so he
may be somewhat tied up. His phone is 920-469-9198 if all else
fails. There's also a Stryke web forum.
http://forum.stryke.com/

Thanks for that, Rusty. I must say, the recent discussions on the
Stryke forum are *not* encouraging to a potential purchaser!

I appreciate the problems if he's a 'one man band' and a new dad, but
at the end of the day I'll need products, not promises............
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Keep us informed on the project and it's results.

I'm planning a project with a single Shiva in a ported box of either 3.6 or
4.8 cu ft. f3 predictions are 25 and 21 Hz, respectively.

Leaning towards the smaller box for WAF nad with room gain it should get to
at least 20 Hz in the room it will be used in.

"Stewart Pinkerton" wrote in message
...
I appreciate the problems if he's a 'one man band' and a new
dad, but
at the end of the day I'll need products, not
promises............

I've been absent from the forum for awhile. It does seem like
John's backed up right now but trying to work out the backlog.
Apparently Stryke will be changing it's name Jan 1st due to
trademark issues. Looks like Acoustic Elegance is the leading
candidate.

FYI, I ordered my AV15s during the first preorder back in Oct
'02. I believe John ordered 500 AV15s from TC Sounds with the
intent to have several hundred left over to keep in stock. TC
Sounds messed up the voice coils (single 4 ohm instead of dual 2
ohm) which delayed shipments until Jan '03. By the time they
shipped John had sold out of all the drivers! Since then Kyle at
www.acoustic-visions.com has been keeping drivers in stock but
they're currently out due to waiting on the new design (dual
spider) but Kyle won't ship outside USA.

If you can handle the wait and get the UK shipping worked out the
AV15 are certainly worth it. If not, go for the Tempest. A 3x2
array of Tempests will be as good as four AV15s or three Tumults


I previously had a Velodyne F1800R and thought it was king. The
IB wipes the floor with it in every way...including cost. I had
planned on performing a level matched DBT but there was no need.
The F1800 now sits behind the computer.

"Stewart Pinkerton" wrote in message
...
I appreciate the problems if he's a 'one man band' and a new
dad, but
at the end of the day I'll need products, not
promises............

I've been absent from the forum for awhile. It does seem like
John's backed up right now but trying to work out the backlog.
Apparently Stryke will be changing it's name Jan 1st due to
trademark issues. Looks like Acoustic Elegance is the leading
candidate.

FYI, I ordered my AV15s during the first preorder back in Oct
'02. I believe John ordered 500 AV15s from TC Sounds with the
intent to have several hundred left over to keep in stock. TC
Sounds messed up the voice coils (single 4 ohm instead of dual 2
ohm) which delayed shipments until Jan '03. By the time they
shipped John had sold out of all the drivers! Since then Kyle at
www.acoustic-visions.com has been keeping drivers in stock but
they're currently out due to waiting on the new design (dual
spider) but Kyle won't ship outside USA.

If you can handle the wait and get the UK shipping worked out the
AV15 are certainly worth it. If not, go for the Tempest. A 3x2
array of Tempests will be as good as four AV15s or three Tumults


I previously had a Velodyne F1800R and thought it was king. The
IB wipes the floor with it in every way...including cost. I had
planned on performing a level matched DBT but there was no need.
The F1800 now sits behind the computer.

"Stewart Pinkerton" wrote in message
...
I appreciate the problems if he's a 'one man band' and a new
dad, but
at the end of the day I'll need products, not
promises............

I've been absent from the forum for awhile. It does seem like
John's backed up right now but trying to work out the backlog.
Apparently Stryke will be changing it's name Jan 1st due to
trademark issues. Looks like Acoustic Elegance is the leading
candidate.

FYI, I ordered my AV15s during the first preorder back in Oct
'02. I believe John ordered 500 AV15s from TC Sounds with the
intent to have several hundred left over to keep in stock. TC
Sounds messed up the voice coils (single 4 ohm instead of dual 2
ohm) which delayed shipments until Jan '03. By the time they
shipped John had sold out of all the drivers! Since then Kyle at
www.acoustic-visions.com has been keeping drivers in stock but
they're currently out due to waiting on the new design (dual
spider) but Kyle won't ship outside USA.

If you can handle the wait and get the UK shipping worked out the
AV15 are certainly worth it. If not, go for the Tempest. A 3x2
array of Tempests will be as good as four AV15s or three Tumults


I previously had a Velodyne F1800R and thought it was king. The
IB wipes the floor with it in every way...including cost. I had
planned on performing a level matched DBT but there was no need.
The F1800 now sits behind the computer.

Stewart Pinkerton wrote:

A slit is certainly an option in my room, although I might have to
write my own modelling program! :-(

Model, why? - ah well, one item of horn modelling software does exist
"out there", I have it somewhere on some harddisk.

Peerless has a 10" sub-bass that looks about right if used in multiples.

If I wanted to feed an elongated opening, then doing it with a line of
units comes to mind.

At the moment, a pair of Adire Tempests looks to be the most likely
solution.

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion. Remember however to also
load the other side of the membrane, at a wild guess and without
checking anything a 5 cubic feet box comes to mind, possibly 7, this to
minimize equal order distortion and to provide some driver protection.

No warranty of any kind, suggested only for those who can freehand the
amount of garlic and curry .....

Stewart Pinkerton | Music is Art - Audio is Engineering

(x) From the unit you make a transition to the slit width, 3" comes to
mind, and then you let the horn expand in one plane only horisontally on
the floor of the loft and then make a 90 degree down turn into the slit.
It is very easy to brace such a flat structure with dividers to ensure
that it is suitably rigid. The slit area should not be smaller than the
cone area, for a "seat of the pants" design guide I'd aim for two to
three times the cone area and use a fairly rapid expansion of the
conical "hornifold".

Kind regards

Peter Larsen

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

Stewart Pinkerton wrote:

A slit is certainly an option in my room, although I might have to
write my own modelling program! :-(

Model, why? - ah well, one item of horn modelling software does exist
"out there", I have it somewhere on some harddisk.

Peerless has a 10" sub-bass that looks about right if used in multiples.

If I wanted to feed an elongated opening, then doing it with a line of
units comes to mind.

At the moment, a pair of Adire Tempests looks to be the most likely
solution.

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion. Remember however to also
load the other side of the membrane, at a wild guess and without
checking anything a 5 cubic feet box comes to mind, possibly 7, this to
minimize equal order distortion and to provide some driver protection.

No warranty of any kind, suggested only for those who can freehand the
amount of garlic and curry .....

Stewart Pinkerton | Music is Art - Audio is Engineering

(x) From the unit you make a transition to the slit width, 3" comes to
mind, and then you let the horn expand in one plane only horisontally on
the floor of the loft and then make a 90 degree down turn into the slit.
It is very easy to brace such a flat structure with dividers to ensure
that it is suitably rigid. The slit area should not be smaller than the
cone area, for a "seat of the pants" design guide I'd aim for two to
three times the cone area and use a fairly rapid expansion of the
conical "hornifold".

Kind regards

Peter Larsen

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

Stewart Pinkerton wrote:

A slit is certainly an option in my room, although I might have to
write my own modelling program! :-(

Model, why? - ah well, one item of horn modelling software does exist
"out there", I have it somewhere on some harddisk.

Peerless has a 10" sub-bass that looks about right if used in multiples.

If I wanted to feed an elongated opening, then doing it with a line of
units comes to mind.

At the moment, a pair of Adire Tempests looks to be the most likely
solution.

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion. Remember however to also
load the other side of the membrane, at a wild guess and without
checking anything a 5 cubic feet box comes to mind, possibly 7, this to
minimize equal order distortion and to provide some driver protection.

No warranty of any kind, suggested only for those who can freehand the
amount of garlic and curry .....

Stewart Pinkerton | Music is Art - Audio is Engineering

(x) From the unit you make a transition to the slit width, 3" comes to
mind, and then you let the horn expand in one plane only horisontally on
the floor of the loft and then make a 90 degree down turn into the slit.
It is very easy to brace such a flat structure with dividers to ensure
that it is suitably rigid. The slit area should not be smaller than the
cone area, for a "seat of the pants" design guide I'd aim for two to
three times the cone area and use a fairly rapid expansion of the
conical "hornifold".

Kind regards

Peter Larsen

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

On Sat, 27 Dec 2003 11:23:09 +0100, Peter Larsen
wrote:

Stewart Pinkerton wrote:

A slit is certainly an option in my room, although I might have to
write my own modelling program! :-(

Model, why? - ah well, one item of horn modelling software does exist
"out there", I have it somewhere on some harddisk.

Peerless has a 10" sub-bass that looks about right if used in multiples.

If I wanted to feed an elongated opening, then doing it with a line of
units comes to mind.

I'm looking at the subwoofer bottom line of Euros per litre! :-)

A dozen Peerless units doesn't cost out so well as a couple of
Tempests, or especially Stryke AV15s.

At the moment, a pair of Adire Tempests looks to be the most likely
solution.

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Remember however to also
load the other side of the membrane, at a wild guess and without
checking anything a 5 cubic feet box comes to mind, possibly 7, this to
minimize equal order distortion and to provide some driver protection.

Yes, this is one of my concerns.

No warranty of any kind, suggested only for those who can freehand the
amount of garlic and curry .....

Always!

Tonight of course, it'll be turkey curry........... :-)

(x) From the unit you make a transition to the slit width, 3" comes to
mind, and then you let the horn expand in one plane only horisontally on
the floor of the loft and then make a 90 degree down turn into the slit.
It is very easy to brace such a flat structure with dividers to ensure
that it is suitably rigid. The slit area should not be smaller than the
cone area, for a "seat of the pants" design guide I'd aim for two to
three times the cone area and use a fairly rapid expansion of the
conical "hornifold".

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Sat, 27 Dec 2003 11:23:09 +0100, Peter Larsen
wrote:

Stewart Pinkerton wrote:

A slit is certainly an option in my room, although I might have to
write my own modelling program! :-(

Model, why? - ah well, one item of horn modelling software does exist
"out there", I have it somewhere on some harddisk.

Peerless has a 10" sub-bass that looks about right if used in multiples.

If I wanted to feed an elongated opening, then doing it with a line of
units comes to mind.

I'm looking at the subwoofer bottom line of Euros per litre! :-)

A dozen Peerless units doesn't cost out so well as a couple of
Tempests, or especially Stryke AV15s.

At the moment, a pair of Adire Tempests looks to be the most likely
solution.

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Remember however to also
load the other side of the membrane, at a wild guess and without
checking anything a 5 cubic feet box comes to mind, possibly 7, this to
minimize equal order distortion and to provide some driver protection.

Yes, this is one of my concerns.

No warranty of any kind, suggested only for those who can freehand the
amount of garlic and curry .....

Always!

Tonight of course, it'll be turkey curry........... :-)

(x) From the unit you make a transition to the slit width, 3" comes to
mind, and then you let the horn expand in one plane only horisontally on
the floor of the loft and then make a 90 degree down turn into the slit.
It is very easy to brace such a flat structure with dividers to ensure
that it is suitably rigid. The slit area should not be smaller than the
cone area, for a "seat of the pants" design guide I'd aim for two to
three times the cone area and use a fairly rapid expansion of the
conical "hornifold".

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Sat, 27 Dec 2003 11:23:09 +0100, Peter Larsen
wrote:

Stewart Pinkerton wrote:

A slit is certainly an option in my room, although I might have to
write my own modelling program! :-(

Model, why? - ah well, one item of horn modelling software does exist
"out there", I have it somewhere on some harddisk.

Peerless has a 10" sub-bass that looks about right if used in multiples.

If I wanted to feed an elongated opening, then doing it with a line of
units comes to mind.

I'm looking at the subwoofer bottom line of Euros per litre! :-)

A dozen Peerless units doesn't cost out so well as a couple of
Tempests, or especially Stryke AV15s.

At the moment, a pair of Adire Tempests looks to be the most likely
solution.

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Remember however to also
load the other side of the membrane, at a wild guess and without
checking anything a 5 cubic feet box comes to mind, possibly 7, this to
minimize equal order distortion and to provide some driver protection.

Yes, this is one of my concerns.

No warranty of any kind, suggested only for those who can freehand the
amount of garlic and curry .....

Always!

Tonight of course, it'll be turkey curry........... :-)

(x) From the unit you make a transition to the slit width, 3" comes to
mind, and then you let the horn expand in one plane only horisontally on
the floor of the loft and then make a 90 degree down turn into the slit.
It is very easy to brace such a flat structure with dividers to ensure
that it is suitably rigid. The slit area should not be smaller than the
cone area, for a "seat of the pants" design guide I'd aim for two to
three times the cone area and use a fairly rapid expansion of the
conical "hornifold".

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Stewart Pinkerton wrote:

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Expansion prevents too many parallel walls. If you go for the suggested
model with 3" as the smallest dimension of manifold and slit then you
have to make some kind of a transition from the 13" membrane. This is
not about midrange, so doing it the crude way comes to mind, i. e.
starting the manifold with a - tja da dum - 3 by 13 inches or 3 by 10
inches initial area. Even to simply enter the room with the membrane
area will then require expansion.

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

Ah, going to the room corner is a mitigating factor because it allows
you to use the corner as an extension of the conical horn in the
manifold. And with an "L" shaped slit in the ceiling in the corner it
ought to be possible to make it a lot more compact than I had first
thougth.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................

Minimizing distortion and getting an optimum coupling to the room are
worthwhile aims and probably pre-requites for a system that is supposed
to be able to play well in trippel forte as well as in the way more
difficult trippel piano.

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Stewart Pinkerton | Music is Art - Audio is Engineering


Kind regards

Peter Larsen

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

Stewart Pinkerton wrote:

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Expansion prevents too many parallel walls. If you go for the suggested
model with 3" as the smallest dimension of manifold and slit then you
have to make some kind of a transition from the 13" membrane. This is
not about midrange, so doing it the crude way comes to mind, i. e.
starting the manifold with a - tja da dum - 3 by 13 inches or 3 by 10
inches initial area. Even to simply enter the room with the membrane
area will then require expansion.

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

Ah, going to the room corner is a mitigating factor because it allows
you to use the corner as an extension of the conical horn in the
manifold. And with an "L" shaped slit in the ceiling in the corner it
ought to be possible to make it a lot more compact than I had first
thougth.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................

Minimizing distortion and getting an optimum coupling to the room are
worthwhile aims and probably pre-requites for a system that is supposed
to be able to play well in trippel forte as well as in the way more
difficult trippel piano.

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Stewart Pinkerton | Music is Art - Audio is Engineering


Kind regards

Peter Larsen

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

Stewart Pinkerton wrote:

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Expansion prevents too many parallel walls. If you go for the suggested
model with 3" as the smallest dimension of manifold and slit then you
have to make some kind of a transition from the 13" membrane. This is
not about midrange, so doing it the crude way comes to mind, i. e.
starting the manifold with a - tja da dum - 3 by 13 inches or 3 by 10
inches initial area. Even to simply enter the room with the membrane
area will then require expansion.

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

Ah, going to the room corner is a mitigating factor because it allows
you to use the corner as an extension of the conical horn in the
manifold. And with an "L" shaped slit in the ceiling in the corner it
ought to be possible to make it a lot more compact than I had first
thougth.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................

Minimizing distortion and getting an optimum coupling to the room are
worthwhile aims and probably pre-requites for a system that is supposed
to be able to play well in trippel forte as well as in the way more
difficult trippel piano.

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Stewart Pinkerton | Music is Art - Audio is Engineering


Kind regards

Peter Larsen

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

"Stewart Pinkerton" wrote in message
...
OTOH, since I'm staying well below Xmax for normal operation,
perhaps
this is all a little paranoid! OTGH, design time costs me
nothing, so
why not do it right first time..................

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

Also, if you didn't see there's a set a photos on the first
gallery page of the IB site where someone built a cloth covered
grill for a white ceiling. I would have made it flush with the
ceiling...maybe something else to think about for wife
acceptance.

"Stewart Pinkerton" wrote in message
...
OTOH, since I'm staying well below Xmax for normal operation,
perhaps
this is all a little paranoid! OTGH, design time costs me
nothing, so
why not do it right first time..................

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

Also, if you didn't see there's a set a photos on the first
gallery page of the IB site where someone built a cloth covered
grill for a white ceiling. I would have made it flush with the
ceiling...maybe something else to think about for wife
acceptance.

"Stewart Pinkerton" wrote in message
...
OTOH, since I'm staying well below Xmax for normal operation,
perhaps
this is all a little paranoid! OTGH, design time costs me
nothing, so
why not do it right first time..................

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

Also, if you didn't see there's a set a photos on the first
gallery page of the IB site where someone built a cloth covered
grill for a white ceiling. I would have made it flush with the
ceiling...maybe something else to think about for wife
acceptance.

Rusty Boudreaux wrote:

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

The nice thing about the slit is that one can install lights in it and a
bit of screen in front of it and pretend that it is mood lights for the
room.


--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

Rusty Boudreaux wrote:

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

The nice thing about the slit is that one can install lights in it and a
bit of screen in front of it and pretend that it is mood lights for the
room.


--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

Rusty Boudreaux wrote:

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

The nice thing about the slit is that one can install lights in it and a
bit of screen in front of it and pretend that it is mood lights for the
room.


--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

On Sat, 27 Dec 2003 22:18:15 -0600, "Rusty Boudreaux"
wrote:

"Stewart Pinkerton" wrote in message
...
OTOH, since I'm staying well below Xmax for normal operation,
perhaps
this is all a little paranoid! OTGH, design time costs me
nothing, so
why not do it right first time..................

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

Nope, it has 13" thick brick/block walls, only the ceiling is
conventional wooden joists and plasterboard (sheetrock to you).

Also, if you didn't see there's a set a photos on the first
gallery page of the IB site where someone built a cloth covered
grill for a white ceiling. I would have made it flush with the
ceiling...maybe something else to think about for wife
acceptance.

Yes, I've seen that shot, and I might get away with something similar
to that, or two or three of the HVAC vents someone else used. My
concern with the cloth approach is that a Vd of 5 litres at 20-50Hz is
likely to cause some pretty visible flapping!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Sat, 27 Dec 2003 22:18:15 -0600, "Rusty Boudreaux"
wrote:

"Stewart Pinkerton" wrote in message
...
OTOH, since I'm staying well below Xmax for normal operation,
perhaps
this is all a little paranoid! OTGH, design time costs me
nothing, so
why not do it right first time..................

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

Nope, it has 13" thick brick/block walls, only the ceiling is
conventional wooden joists and plasterboard (sheetrock to you).

Also, if you didn't see there's a set a photos on the first
gallery page of the IB site where someone built a cloth covered
grill for a white ceiling. I would have made it flush with the
ceiling...maybe something else to think about for wife
acceptance.

Yes, I've seen that shot, and I might get away with something similar
to that, or two or three of the HVAC vents someone else used. My
concern with the cloth approach is that a Vd of 5 litres at 20-50Hz is
likely to cause some pretty visible flapping!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Sat, 27 Dec 2003 22:18:15 -0600, "Rusty Boudreaux"
wrote:

"Stewart Pinkerton" wrote in message
...
OTOH, since I'm staying well below Xmax for normal operation,
perhaps
this is all a little paranoid! OTGH, design time costs me
nothing, so
why not do it right first time..................

If your house is drywall/sheetrock just go for it. Cheap and
easy to patch

Nope, it has 13" thick brick/block walls, only the ceiling is
conventional wooden joists and plasterboard (sheetrock to you).

Also, if you didn't see there's a set a photos on the first
gallery page of the IB site where someone built a cloth covered
grill for a white ceiling. I would have made it flush with the
ceiling...maybe something else to think about for wife
acceptance.

Yes, I've seen that shot, and I might get away with something similar
to that, or two or three of the HVAC vents someone else used. My
concern with the cloth approach is that a Vd of 5 litres at 20-50Hz is
likely to cause some pretty visible flapping!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Sat, 27 Dec 2003 19:03:59 +0100, Peter Larsen
wrote:

Stewart Pinkerton wrote:

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Expansion prevents too many parallel walls.

Which are completely irrelevant in a sub. In fact, a cube is the
*best* enclosure shape for a sealed-box sub, as it has maxium wall
rigidity for any given volume and material.

If you go for the suggested
model with 3" as the smallest dimension of manifold and slit then you
have to make some kind of a transition from the 13" membrane. This is
not about midrange, so doing it the crude way comes to mind, i. e.
starting the manifold with a - tja da dum - 3 by 13 inches or 3 by 10
inches initial area. Even to simply enter the room with the membrane
area will then require expansion.

You can however transit directly from a 13" circle to a 3" x 44" slit
(to maintain equal areas) without any expansion whatever, simply by
making a double-tapered box, so why bother with an intermediate
constriction? BTW, such a manifold box will have minimal resonance
problems due to the heavy taper of the box sides, so it's also a
mechanically sound design not requiring heroic wall thickness.

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

Ah, going to the room corner is a mitigating factor because it allows
you to use the corner as an extension of the conical horn in the
manifold. And with an "L" shaped slit in the ceiling in the corner it
ought to be possible to make it a lot more compact than I had first
thougth.

I currently favour a 10" square aperture, which is reasonably discreet
and gives a 3:1 area compression from a pair of 15" drivers. Might
have to be a foot square to allow for grille solidity.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................

Minimizing distortion and getting an optimum coupling to the room are
worthwhile aims and probably pre-requites for a system that is supposed
to be able to play well in trippel forte as well as in the way more
difficult trippel piano.

Indeed so, hence lots of thinking before firing up the jigsaw!

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Aside from the basic expansion into 1/8 space given by corner
mounting, I have no intention of getting into the multitudinous
problems of horns!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Sat, 27 Dec 2003 19:03:59 +0100, Peter Larsen
wrote:

Stewart Pinkerton wrote:

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Expansion prevents too many parallel walls.

Which are completely irrelevant in a sub. In fact, a cube is the
*best* enclosure shape for a sealed-box sub, as it has maxium wall
rigidity for any given volume and material.

If you go for the suggested
model with 3" as the smallest dimension of manifold and slit then you
have to make some kind of a transition from the 13" membrane. This is
not about midrange, so doing it the crude way comes to mind, i. e.
starting the manifold with a - tja da dum - 3 by 13 inches or 3 by 10
inches initial area. Even to simply enter the room with the membrane
area will then require expansion.

You can however transit directly from a 13" circle to a 3" x 44" slit
(to maintain equal areas) without any expansion whatever, simply by
making a double-tapered box, so why bother with an intermediate
constriction? BTW, such a manifold box will have minimal resonance
problems due to the heavy taper of the box sides, so it's also a
mechanically sound design not requiring heroic wall thickness.

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

Ah, going to the room corner is a mitigating factor because it allows
you to use the corner as an extension of the conical horn in the
manifold. And with an "L" shaped slit in the ceiling in the corner it
ought to be possible to make it a lot more compact than I had first
thougth.

I currently favour a 10" square aperture, which is reasonably discreet
and gives a 3:1 area compression from a pair of 15" drivers. Might
have to be a foot square to allow for grille solidity.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................

Minimizing distortion and getting an optimum coupling to the room are
worthwhile aims and probably pre-requites for a system that is supposed
to be able to play well in trippel forte as well as in the way more
difficult trippel piano.

Indeed so, hence lots of thinking before firing up the jigsaw!

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Aside from the basic expansion into 1/8 space given by corner
mounting, I have no intention of getting into the multitudinous
problems of horns!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Sat, 27 Dec 2003 19:03:59 +0100, Peter Larsen
wrote:

Stewart Pinkerton wrote:

Freehand a manifold, i. e. a conical horn (x), as the adapter between
the unit and the slit. Conical horns are wonderfully simple: the faster
they expand the higher their range. They also have the advantage of low
distortion due to the fast initial expansion.

This doesn't make sense to me, as there is no horn expansion if the
slit is the same area as the driver Sd, as you suggest below.

Expansion prevents too many parallel walls.

Which are completely irrelevant in a sub. In fact, a cube is the
*best* enclosure shape for a sealed-box sub, as it has maxium wall
rigidity for any given volume and material.

If you go for the suggested
model with 3" as the smallest dimension of manifold and slit then you
have to make some kind of a transition from the 13" membrane. This is
not about midrange, so doing it the crude way comes to mind, i. e.
starting the manifold with a - tja da dum - 3 by 13 inches or 3 by 10
inches initial area. Even to simply enter the room with the membrane
area will then require expansion.

You can however transit directly from a 13" circle to a 3" x 44" slit
(to maintain equal areas) without any expansion whatever, simply by
making a double-tapered box, so why bother with an intermediate
constriction? BTW, such a manifold box will have minimal resonance
problems due to the heavy taper of the box sides, so it's also a
mechanically sound design not requiring heroic wall thickness.

I'm still looking at a manifold exit area of 0.3-0.5 the Sd of the two
drivers (for improved WAF....) , expanding into the 1/8 space of the
room corner. As you say, there may need to be a back box to equalise
cone loading.

Ah, going to the room corner is a mitigating factor because it allows
you to use the corner as an extension of the conical horn in the
manifold. And with an "L" shaped slit in the ceiling in the corner it
ought to be possible to make it a lot more compact than I had first
thougth.

I currently favour a 10" square aperture, which is reasonably discreet
and gives a 3:1 area compression from a pair of 15" drivers. Might
have to be a foot square to allow for grille solidity.

OTOH, since I'm staying well below Xmax for normal operation, perhaps
this is all a little paranoid! OTGH, design time costs me nothing, so
why not do it right first time..................

Minimizing distortion and getting an optimum coupling to the room are
worthwhile aims and probably pre-requites for a system that is supposed
to be able to play well in trippel forte as well as in the way more
difficult trippel piano.

Indeed so, hence lots of thinking before firing up the jigsaw!

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Aside from the basic expansion into 1/8 space given by corner
mounting, I have no intention of getting into the multitudinous
problems of horns!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Sorry guys - I missed the start of this thread, and I've been watching ever
since to find out just what is an "IB"?

Tony (remove the "_" to reply by email)

Sorry guys - I missed the start of this thread, and I've been watching ever
since to find out just what is an "IB"?

Tony (remove the "_" to reply by email)

Sorry guys - I missed the start of this thread, and I've been watching ever
since to find out just what is an "IB"?

Tony (remove the "_" to reply by email)

On Mon, 29 Dec 2003 01:40:54 +1000, Tony Roe
wrote:

Sorry guys - I missed the start of this thread, and I've been watching ever
since to find out just what is an "IB"?

Infinite Baffle. Eithee a *very* large plate to which the drivers are
attached, or more reasonably, fitting the drivers in the wall between
two rooms.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Mon, 29 Dec 2003 01:40:54 +1000, Tony Roe
wrote:

Sorry guys - I missed the start of this thread, and I've been watching ever
since to find out just what is an "IB"?

Infinite Baffle. Eithee a *very* large plate to which the drivers are
attached, or more reasonably, fitting the drivers in the wall between
two rooms.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Mon, 29 Dec 2003 01:40:54 +1000, Tony Roe
wrote:

Sorry guys - I missed the start of this thread, and I've been watching ever
since to find out just what is an "IB"?

Infinite Baffle. Eithee a *very* large plate to which the drivers are
attached, or more reasonably, fitting the drivers in the wall between
two rooms.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Stewart Pinkerton wrote:

Expansion prevents too many parallel walls.

Which are completely irrelevant in a sub ....

We are not in the sub, we are on the way into the living room to stir
the coffee.

You can however transit directly from a 13" circle to a 3" x 44" slit
(to maintain equal areas) without any expansion whatever, simply by
making a double-tapered box, so why bother with an intermediate
constriction?

To avoid making the double tapered box ...

I currently favour a 10" square aperture, which is reasonably discreet
and gives a 3:1 area compression from a pair of 15" drivers. Might
have to be a foot square to allow for grille solidity.

Erm, what happened to the 3# by 44# slit above .... to have the driver
chassis vertical wold of course require a 90 degree downwards bend to
enter the room.

Indeed so, hence lots of thinking before firing up the jigsaw!

It could well be an advantage when later negotiating other building
permits to have this project well planned and neatly executed and "right
at the first go".

[re-iterating]

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Aside from the basic expansion into 1/8 space given by corner
mounting, I have no intention of getting into the multitudinous
problems of horns!

The corner *is* a conical horn, whether you want it to so be or not, and
seen like that it could be that hornresp can suggest the volume for the
rear load chamber as well as give a hint about what kind of frequency
response to expect below Fs. Frankly however this is rapidly leaving my
field of knowhow, my gut feeling remains however: horn it.

Stewart Pinkerton | Music is Art - Audio is Engineering

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

Stewart Pinkerton wrote:

Expansion prevents too many parallel walls.

Which are completely irrelevant in a sub ....

We are not in the sub, we are on the way into the living room to stir
the coffee.

You can however transit directly from a 13" circle to a 3" x 44" slit
(to maintain equal areas) without any expansion whatever, simply by
making a double-tapered box, so why bother with an intermediate
constriction?

To avoid making the double tapered box ...

I currently favour a 10" square aperture, which is reasonably discreet
and gives a 3:1 area compression from a pair of 15" drivers. Might
have to be a foot square to allow for grille solidity.

Erm, what happened to the 3# by 44# slit above .... to have the driver
chassis vertical wold of course require a 90 degree downwards bend to
enter the room.

Indeed so, hence lots of thinking before firing up the jigsaw!

It could well be an advantage when later negotiating other building
permits to have this project well planned and neatly executed and "right
at the first go".

[re-iterating]

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Aside from the basic expansion into 1/8 space given by corner
mounting, I have no intention of getting into the multitudinous
problems of horns!

The corner *is* a conical horn, whether you want it to so be or not, and
seen like that it could be that hornresp can suggest the volume for the
rear load chamber as well as give a hint about what kind of frequency
response to expect below Fs. Frankly however this is rapidly leaving my
field of knowhow, my gut feeling remains however: horn it.

Stewart Pinkerton | Music is Art - Audio is Engineering

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

Stewart Pinkerton wrote:

Expansion prevents too many parallel walls.

Which are completely irrelevant in a sub ....

We are not in the sub, we are on the way into the living room to stir
the coffee.

You can however transit directly from a 13" circle to a 3" x 44" slit
(to maintain equal areas) without any expansion whatever, simply by
making a double-tapered box, so why bother with an intermediate
constriction?

To avoid making the double tapered box ...

I currently favour a 10" square aperture, which is reasonably discreet
and gives a 3:1 area compression from a pair of 15" drivers. Might
have to be a foot square to allow for grille solidity.

Erm, what happened to the 3# by 44# slit above .... to have the driver
chassis vertical wold of course require a 90 degree downwards bend to
enter the room.

Indeed so, hence lots of thinking before firing up the jigsaw!

It could well be an advantage when later negotiating other building
permits to have this project well planned and neatly executed and "right
at the first go".

[re-iterating]

Try searching for hornresp ... it may be possible to express this as a
conical horn it can understand, in which case it could be helpful to try
modelling in it.

Aside from the basic expansion into 1/8 space given by corner
mounting, I have no intention of getting into the multitudinous
problems of horns!

The corner *is* a conical horn, whether you want it to so be or not, and
seen like that it could be that hornresp can suggest the volume for the
rear load chamber as well as give a hint about what kind of frequency
response to expect below Fs. Frankly however this is rapidly leaving my
field of knowhow, my gut feeling remains however: horn it.

Stewart Pinkerton | Music is Art - Audio is Engineering

--
************************************************** *************
* \\\\\\\ Quality Ascii handcrafted by Peter Larsen /////// *
* \\\\\\\ My site is at: http://www.muyiovatki.dk /////// *
************************************************** *******

"Peter Larsen" wrote in message
...
The corner *is* a conical horn, whether you want it to so be or
not, and
seen like that it could be that hornresp can suggest the volume
for the
rear load chamber as well as give a hint about what kind of
frequency
response to expect below Fs. Frankly however this is rapidly
leaving my
field of knowhow, my gut feeling remains however: horn it.

I've kinda lost track of this thread. If we're still talking
about an infinite baffled subwoofer with drivers mounted in a
manifold ported into the listening area then it most certainly is
NOT a horn. Even at a high frequency of 100Hz the manifold
dimensions are less than 1/10 of a wavelength. The manifold will
act as a pressure source at the opening into the listening area.
Of the hundreds of people who have built an IB sub and posted
their results on the IB forum no one has ever used anything other
than a square/rectangle opening.

Just construct a box from 3/4" plywood, put drivers on opposite
sides, make the box as short as possible (drivers as close as
possible to listening room) and use as large an opening as
possible. Pretty simple, eh?

For four 15" drivers using a cube of 18" or so is ideal. Even
with two drivers still use a cube as you can easily add two more
drivers later if you want more output.

Below is a perfect example with response measurements. Just use
four drivers instead of twelve.
http://t-3.cc/users/audioworx/page13-12Shiva1.html

Or you can build a manifold the same width as your rafters and
put two drivers on each side.
http://t-3.cc/users/audioworx/page25Mini-Me1.html

"Peter Larsen" wrote in message
...
The corner *is* a conical horn, whether you want it to so be or
not, and
seen like that it could be that hornresp can suggest the volume
for the
rear load chamber as well as give a hint about what kind of
frequency
response to expect below Fs. Frankly however this is rapidly
leaving my
field of knowhow, my gut feeling remains however: horn it.

I've kinda lost track of this thread. If we're still talking
about an infinite baffled subwoofer with drivers mounted in a
manifold ported into the listening area then it most certainly is
NOT a horn. Even at a high frequency of 100Hz the manifold
dimensions are less than 1/10 of a wavelength. The manifold will
act as a pressure source at the opening into the listening area.
Of the hundreds of people who have built an IB sub and posted
their results on the IB forum no one has ever used anything other
than a square/rectangle opening.

Just construct a box from 3/4" plywood, put drivers on opposite
sides, make the box as short as possible (drivers as close as
possible to listening room) and use as large an opening as
possible. Pretty simple, eh?

For four 15" drivers using a cube of 18" or so is ideal. Even
with two drivers still use a cube as you can easily add two more
drivers later if you want more output.

Below is a perfect example with response measurements. Just use
four drivers instead of twelve.
http://t-3.cc/users/audioworx/page13-12Shiva1.html

Or you can build a manifold the same width as your rafters and
put two drivers on each side.
http://t-3.cc/users/audioworx/page25Mini-Me1.html